This was basically the whole interview in one question.
Start by clarifying requirements (scale, security level, use cases) and then walk through the PKI design in layers: hierarchy, issuance, revocation, key storage, lifecycle, and validation. Emphasize trade-offs and how each component integrates with the others, and conclude with how you would monitor and operate the system.
Pro tip: Show awareness of real-world constraints like certificate expiration and revocation propagation delays, and propose mitigations such as short-lived certificates and OCSP stapling. Mention automation and infrastructure-as-code for managing PKI at scale, which is critical for a company like Coupang.
Ask about scale (number of certificates, issuance rate), security requirements (key sizes, algorithms), compliance (e.g., PCI DSS), and existing infrastructure. This ensures your design is tailored and demonstrates you think before designing.
Propose a root CA (offline, highly secured) and intermediate CAs for different purposes (e.g., internal services, external-facing). Explain the rationale for separation to limit blast radius and enable flexible policy enforcement.
Describe the enrollment process (e.g., CSR generation, identity validation), the signing workflow (possibly automated via API), and how to handle certificate templates and profiles. Mention protocols like ACME for automation.
Compare CRL vs OCSP, discuss trade-offs (latency, scalability), and propose a hybrid approach with OCSP stapling. Outline certificate lifecycle: issuance, renewal, revocation, and expiration, with automation for renewal and monitoring for expirations.
Explain key storage options (HSM, TPM, secure enclaves) for CAs and clients, and how clients validate the trust chain (certificate path validation, checking revocation, hostname verification). Mention pinning and trust stores.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.